arasharchor / dlsg

Deep Learning Study Group

DLSG

Deep Learning Study Group

Members - Antoine Toisoul, Giacomo Tarroni, Mohamed Elmikaty, Enzo Ferrante, Amir Alansary

This is a group of researchers study deep learning for computer vision, medical imaging, and computer graphics. Here we define the scheduled dates and material for the regular (weekly) meetings. The main goal is to discuss and exchange ideas and information about a specific subject from deep learning related to our research.

---

Convolutional Neural Networks Basics

For a better understanding and applying deep learning in computer vision, we follow Stanford course in the first weeks below:

Week-01

• Intro to Computer Vision, historical context [slides] [video]
• Image classification, k-nearest neighbor, and linear classification [slides] [video]
  • Image classification [notes]
  • Linear classification [notes]

Week-02

• Optimization, stochastic gradient descent [notes] [video]

Week-03

• Python/numpy tutorial [link]
• Assignment #1 (kNN/SVM/Softmax) [link]
• [Extra] Style Guide for Python Code [link]
• [Extra] Kevin Keraundren's Tutorial for Python and Medical Image Analysis [link]

Week-04

• Backpropagation and Introduction to neural networks [notes] [video]
• Neural Networks Part 1 [notes]
• [Extra] Backpropagation MIT lecture [video]

Week-05

• Neural network part-II -- setting up the data and the model [notes] [video]
• Neural network part-III -- learning the network [notes] [video]

Week-06

• Assignment #1 (2-Layer Net) [link]
• Neural network case study [link]

Week-07

• Assignment #2 (Fully-connected Neural Network/Batch Normalization/Dropout) [link]
• [Extra] Batch normalization backpropagation [link1] [link2]

Week-08

• Convolutional Neural Networks: architectures, convolution / pooling layers [notes]
• Assignment #2 (ConvNet on CIFAR-10) [link]

---

Useful Links for Deep Learning

Important Links

• Awesome deep vision
• Awesome computer vision
• Awesome machine learning
• Deeplearning.net
• Deep learning nature article
• CVIT summer school resources

Libraries

• Software links
• Pyimagesearch

Tutorials

• Theano
• Lasagne
• Caffe
• Tensorflow
• Keras Deep Learning Tutorial for Kaggle Ultrasound Nerve Segmentation competition, using Keras

Courses

• Stanford
• Hinton
• Nando de Freitas
• Hugo Larochelle
• Udacity Deep Learning Course
• Bargava from Cisco systems

Videos

• Deep Learning Summer School (DLSS), Montreal 2016
• Hugo Larochelle's Youtube
• Nando de Freitas's Youtube
• CS231 Winter 2016 All Youtube

Textbooks

• Ian Goodfellow, Yoshua Bengio and Aaron Courville textbook
• Microsoft textbook
• Michael Nielsen's online textbook

Blogs

• Hacker News for Data Science
• i am trusk
• Colin Morris